Machine Learning Enables Highly Accurate Predictions of Photophysical Properties of Organic Fluorescent Materials: Emission Wavelengths and Quantum Yields

Abstract

<div> <p>The predictions of photophysical parameters are of crucial practical importance for the development of functional organic fluorescent materials, whereas the expense of quantum mechanical calculations and the relatively low universality of QSAR models have challenged the task. New avenues opened up by machine learning (ML), we establish a database of solvated organic fluorescent dyes and develop highly efficient ML models for the predictions of maximum emission/absorption wavelength and photoluminescence quantum yields, providing a reliable and efficient approach to high-throughput screenings. Various combinations of ML algorithms and molecular fingerprints were investigated. For emission wavelengths, TD-DFT accuracy was achieved under real-world conditions. Reliable identification of strong fluorescent materials was also demonstrated. We show that the easily obtainable consensus fingerprint inputs combined with proper ML algorithms enables efficient re-training based on additional datapoints whereby systematic improvements of our ML models can be achieved. </p></div>

Machine Learning for Electronically Excited States of Molecules

“…Different types of ML models, such as KRR and NNs, were further compared to predict emission and absorption wavelengths and luminescence quantum yields of organic dyes that show fluorescence after excitation. Different fingerprints were compared generated from an external software and data was obtained from experiments . The tested model emphasized the possibility to combine experimental data with ML algorithms to enable large scale screening and the design of novel materials and compounds.…”